使用点绘制Pandas数据帧中的条形图

Question

I want to plot a horizontal barplot from a Pandas dataframe but have no idea how to begin.

My data looks like this

          max  min  point1  point2
Series 1   50   10      40      30
Series 2   60   20      50      40

Couldn't help myself but to draw something with paint. I want to end up with something like this:

The colors are not important. Here is the dataframe:

import pandas as pd

data = pd.DataFrame(dict(min=[10, 20],
                          max=[50, 60],
                          point1=[40, 50],
                          point2=[30, 40]),
                          index=["Series 1", "Series 2"])

Answer 1

Here is a plot that closely resembles to picture from the question. It's produced by matplotlib.pyplot .

import pandas as pd
import matplotlib.pyplot as plt


data = pd.DataFrame(dict(min=[10, 20],
                          max=[50, 60],
                          point1=[40, 50],
                          point2=[30, 40]),
                          index=["Series 1", "Series 2"])

plt.barh(range(len(data)), data["max"]-data["min"], height=0.3, left=data["min"])

plt.plot(data["point1"], range(len(data)), linestyle="", markersize=10, 
         marker="o", color="#ff6600", label="Point 1", markeredgecolor="k")
plt.plot(data["point2"], range(len(data)), linestyle="", markersize=10, 
         marker="o", color="#aa0000", label="Point 2", markeredgecolor="k")
plt.yticks(range(len(data)),data.index)

plt.ylim(-1,3)
plt.xlim(0,80)
plt.legend()
plt.show()

Answer 2

You can do any drawing you need with matplotlib .

You can specify the bottoms / tops of horizontal barplots in matplotlib like so (using your dataframe as a reference):

import numpy as np
import matplotlib.pyplot as plt

# make a bar floating in space
index_labels = data.index
index = np.arange(index_labels)
right_edge = data['max']
left_edge = data['min']
plt.barh(index, right_edge, left=left_edge)
plt.yticks(index, index_labels)

That should generate a horizontal bar plot where you specify the starting and ending points of the bars. You can reference the docs here: http://matplotlib.org/api/pyplot_api.html#matplotlib.pyplot.barh and see an example here: http://matplotlib.org/1.2.1/examples/pylab_examples/barh_demo.html .

Answer 3

I do not know your data, but pandas can plot boxplots from the plane data by aggregating min, max and percentiles automatically. One example, close to yours might be this:

import pandas as pd

pd.DataFrame({
    "Series 2":[20,25,20,30,45,60], 
    "Series 1":[10,12,20,40,45,50]
}).plot.box(vert=False)

Answer 4

If you realy need a hand-crafted plots width dots and rectangles you may do this like here:

from matplotlib import pyplot as plt
from matplotlib.patches import Rectangle, Circle

fig = plt.figure()
ax = fig.add_subplot(111, xlim=(0, 70), ylim=(0, 70))
ax.add_artist(Rectangle(xy=(10, 49), color="g", width=40, height=2))
ax.add_artist(Circle(xy=(30, 50), color="orange", radius=1))
ax.add_artist(Circle(xy=(40, 50), color="red", radius=1))
ax.add_artist(Rectangle(xy=(20, 19), color="b", width=40, height=2))
ax.add_artist(Circle(xy=(40, 20), color="orange", radius=1))
ax.add_artist(Circle(xy=(50, 20), color="red", radius=1))
plt.show()